Mistral AI has announced three steps at once under the heading of AI sovereignty: strengthening the regional control of inference, expanding access to third-party open models within its own infrastructure, and bringing institutions together to secure long-term compute capacity in Europe.

The company's framing is direct: every enterprise and country, it argues, must control the models it uses, choose where the intelligence runs, control the compute capacity to scale it, and retain the compounding value that follows.

Where inference runs

Most of Mistral's customers run its models inside their own data centres and cloud environments. As AI becomes more deeply embedded in production, they need confidence that the underlying capacity will remain available and resilient.

The company answers that with two products:

  • Mistral Regional Endpoints (generally available): customers choose whether inference runs in Europe or the US, aligning inference location with data-residency, regulatory and latency requirements. Inference and associated processing take place in the selected region, subject to limited, safeguarded transfers to sub-processors that may occur outside it.
  • Mistral Priority Tier (public preview): committed service levels for mission-critical workloads, including custom rate limits, backed by an uptime SLA.

Mistral says it is the only European AI lab to offer both: a choice of processing region and a committed, SLA-backed service level.

Model choice as part of sovereignty

Mistral's argument is that controlling where AI runs is only one part of sovereignty. Customers also need control over which intelligence they run on that infrastructure.

The reasoning is technical: today's AI systems are no longer built on a single model but on an ensemble of capabilities — extended reasoning, high-volume production, or work shaped around a company's own data. That is why the company is expanding access to third-party open models within its infrastructure.

The 2030 target

The third step is the longest-dated. Mistral is convening enterprises and institutions in a coalition to secure long-term commitments for compute capacity in Europe, with plans to build up to one gigawatt of capacity by 2030.

The significance of that figure lies in who owns the capacity. European organisations today depend largely on infrastructure owned and operated outside Europe; capacity built on the continent reduces that dependency. Whether the commitments materialise, and whether the gigawatt target holds, is a promise measurable only at the end of the next four years.

Why sovereignty is on the agenda

In Europe, the AI sovereignty debate is not merely a marketing heading. The General Data Protection Regulation and sector-specific rules require many organisations to document where their data is processed. For them, being able to choose which continent inference runs on is a requirement rather than a preference.

The capacity side, however, is not solved by regulation. Even when an organisation chooses to run inference in Europe, the owner and operator of that capacity is often outside it. Mistral's coalition step targets exactly that gap — and that is why it is the hardest of the three: writing a regulation is faster than building a data centre.